The presence of microbial contamination in clinical specimens is conventionally determined by culturing the specimens in the presence of nutrients and detecting microbial activity through changes in the specimen or in the atmosphere over the specimen after a period of time. For example, in U.S. Pat. No. 4,182,656 to Ahnell et al the sample is placed in a container with a culture medium comprising a carbon 13 labelled fermentable substrate. After sealing the container and subjecting the specimen to conditions conducive to biological activity, the ratio of carbon 13 to carbon 12 in the gaseous atmosphere over the specimen is determined and compared with the initial ratio. In U.S. Pat. No. 4,152,213, a method is claimed by which the presence of oxygen consuming bacteria in a specimen is determined in a sealed container by detecting a reduction in the amount of oxygen in the atmosphere over the specimen through monitoring the pressure of the gas in the container. U.S. Pat. No. 4,073,691 provides a method for determining the presence of biologically active agents, including bacteria, in a sealed container containing a culture medium by measuring changes in the character of the gaseous atmosphere over the specimen after a period of time. A method for non-invasive detection of CO.sub.2 changes in the gaseous atmosphere is taught by Suppman et al, as disclosed in EPO application Ser. No. 83108468.6, published Apr. 4, 1984. The methods and apparatus described in these and other publications all require either a radiometric method or the invasion of the sealed container to measure changes in the gaseous atmosphere after culturing or require special materials that permit infra-red light to pass.
Other known methods for measuring microbial contamination of specimens, particularly blood cultures, include measuring minute changes in temperature, pH, turbidity, color, bioluminescence, and impedance. Generally, these methods determine microbial contamination by detecting bacterial metabolic byproducts. Microbial contamination may also be assessed by subculturing and/or staining. Of these methods, only impedance, radiometry and infra-red spectrometry provide the possibility of automated processing of clinical specimens. And except for impedance and infra-red measurements, these procedures also require entering the container in order to make a measurement on the liquid specimen or the gaseous atmosphere over the specimen. In addition to the likelihood of contamination and creating the likelihood of altering the constituency of the atmosphere over the specimen each time a determination is made, these methods do not permit taking measurements continuously or repeatedly over short time intervals for an extended period of time. This is a significant disadvantage as the rate of growth of contaminating organisms differs depending on the organism and the number of organisms in the original sample, such that it cannot be predicted when detectable changes in the atmosphere or fluid sample will be presented by a contaminated specimen. In a related problem, when contamination is determined by pH changes in the liquid sample, various metabolic products will affect the pH of the sample differently. For example, the production of ammonia will raise the pH while the production of CO.sub.2 will lower it. Different growth rates of different contaminating organisms could result in a pH increase at one time and a decrease at another time, which would not be detected if the pH is measured at widely spaced intervals. Another source of error when detecting changes by pH measurement in whole blood samples, particularly when an indicator dye is the means for pH determination, is the likelihood that the dye appearance can be affected or obscured by the presence of blood cells. Colorimetric indicators can only be effectively used if errors induced by the nature of the specimen can be prevented from influencing the appearance of the dye.